Climate Dynamics (1995) 12:21-35
( limni¢
Dynamies
© Springer-Verlag 1995
A new snow parameterization
for the M6t o-France climate model
Part I: validation in stand-alone experiments
H. Douville, J.-F. Royer, J.-F. Mahfouf
Mdt6o-France/CNRM, 42 Avenue Coriolis, 31057 Toulouse Cedex, France
Received: 7 November 1994 / Accepted: 30 May 1995
Abstract. Both observational studies and numerical ex-
periments demonstrate the sensitivity of the atmo-
sphere to variations in the extent and mass of snow
cover. There is therefore a need for simple but realistic
snow parameterizations in forecast and climate models.
This study describes a new physically-based snow hy-
drology for use in the Mdtdo-France climate model, to-
gether with the ISBA land-surface scheme. A re-
stricted number of parameters has been added, while
preserving a single surface energy budget. The ageing
process of the snow pack has been introduced through
prognostic equations for snow density and snow albe-
do. Snowmelt computation has been modified over
partially snow-covered and vegetated areas. The new
scheme has been validated against field measurements
in stand-alone simulations forced by observed meteor-
ological conditions. The results show a strong improve-
ment in the model's performance, thereby suggesting
that a simple one-layer snow model is able to repro-
duce the main physical mechanisms governing the
snow pack evolution. Part II of the present study will
concern the validation in a 3-D experiment within the
Mdt6o-France climate model.
1 Introduction
The energy budget of the lower atmosphere is very de-
pendent on the physical properties observed at the
Earth's surface. From this point of view, the snow cov-
er represents a critical component of the climate sys-
tem. The high reflectivity of snow can increase the sur-
face albedo by as much as 60%. The induced radiative
cooling is reinforced by the high thermal emissivity of
the snow pack. The turbulent fluxes are reduced, be-
cause of the increased static stability in the boundary
layer and the decreased roughness length. Fresh snow
acts as a thermal insulator due to its low thermal con-
Correspondence to: H. Douville
ductivity. Snowmelt is a sink for latent heat and re-
leases water to the atmosphere and the ground.
Of all the surface conditions, snow shows the largest
spatial and temporal fluctuations. A single storm can
increase the snow cover extent by the order of 1000 km
in a few days. On longer time scales, snow depth exhi-
bits a strong seasonal cycle, as well as an important in-
terannual variability in the middle latitudes. The physi-
cal properties of snow are also very variable, due to
ageing processes. Snow albedo can vary between 30
and 90%. Snow density increases rapidly with time,
evolving from about 80 kg/m 3 to more than 300 kg/m 3.
Snow thermal properties are very sensitive to the set-
tling of the snow pack.
Early studies of the climatic impact of snow have
been based on observations and have emphasized the
cooling effect of snow on the winter temperatures (De-
wey 1977; Namias 1985). Correlations have been found
between snow cover and geopotential height anomal-
ies, showing that the winter snow pack is determined
by the atmospheric dynamics but also feeds back onto
the atmosphere (Walsh et al. 1982). More sophisticated
statistical approaches have also suggested a high sensi-
tivity of the atmosphere to variations in the extent and
mass of accumulated snow (Walsh et al. 1985). Other
studies have focused on the interannual variability of
the Northern Hemisphere snow cover (Iwasaki 1991;
Gutzler and Rosen 1992; Groisman et al. 1994). They
have shown some relationships between the Eurasian
and North American winter snow packs, as well as a
long term decrease in the springtime Northern Hemis-
phere snow cover extent, which may explain why the
recent global warming is more obvious in spring than
in any other season.
Many recent studies have been with general circula-
tion models (GCMs). They have shown the importance
of snow for weather forecasts as well as for climate si-
mulation (Yeh et al. 1983). The sensitivity of the In-
dian monsoon to the Eurasian snow cover has been
confirmed by numerical experiments (Barnett et al.
1989; Yasunari et al. 1991). The global warming pre-
dicted by climate models in a doubled CO2 atmosphere